In the cloud computing industry, the companies that get mentioned the most are the so-called “big three”: Amazon Web Services (AWS), Microsoft Azure (Azure), Google Cloud Platform (GCP).

Alibaba Cloud is usually ignored in the cloud chatter in mainstream U.S. business media, among Wall Street analysts, and in Silicon Valley media. During Alibaba’s most recent earning call, out of the eight questions analysts asked, only one referenced the cloud unit, and it was in the context of potential effects by the coronavirus (COVID-19). In a deep dive video discussion on GCP’s future by The Information, a premium tech media outlet, only one reporter mentioned Alibaba Cloud offhand in passing.

Is this omission merited, or is it a blindspot?

Earnings and Growth

Let’s look at the earnings and growth rate of Azure, GCP, and Alibaba Cloud to see how they stack up. (Since AWS is the runaway leader in the market at the moment, including it in this comparison seems less meaningful.)

Based on its most recent earning report announced on February 13, 2020, Alibaba Cloud’s previous four quarters of earnings generated a total of 5.1 billion USD in revenue. The year-on-year growth rate compared to the previous four quarters ranged between 62% to 76%.

For Alphabet, which released earnings on its cloud unit (called “Google Cloud” in financial statements) for the first time ever in its most recent report on February 3, 2020, the unit generated 8.9 billion USD. It did not release quarterly results for the previous four quarters, but the annual year-on-year growth rate was 53%. Alphabet does not release numbers on GCP alone, but only as part of the “Google Cloud” category, which also contains other products like GSuite (e.g. Gmail for businesses, domain name management, etc.).

For Microsoft, which calls its cloud unit “Commercial Cloud”, it generated 44.7 billion USD in cloud revenue as announced in its most recent earnings report on January 29, 2020. The year-on-year growth rate of Commercial Cloud for the previous four quarters ranged between 36% to 41%. The year-on-year growth rate of Azure for the previous four quarters ranged between 59% to 73%. Microsoft, like Alphabet, also does not release numbers on Azure alone (except for the growth rate), but only as part of the Commercial Cloud category, which includes a bunch of other products that are delivered via its cloud on a subscription basis, like Office 365 Suite, Teams, probably Skype (since its part of the office productivity offering), and other services.

44.7 > 8.9 > 5.1. 2nd, 3rd, 4th place. Case settled, right?

Not All Clouds Are Created Equal

It’s evident that the way these tech companies report on their cloud businesses is wildly inconsistent and convoluted, which makes assessing how a cloud computing platform is doing relative to its peers difficult.

However, for an engineer or a developer, what makes a cloud is clear. In plain language: it’s a set of compute, storage, and networking resources, plus infrastructure-level technology services (like databases, container orchestration, big data tools, machine learning frameworks) that are always available to be rented out to build applications.

That means: Azure is a cloud, “Commercial Cloud” is not. GCP is a cloud, “Google Cloud” is not. And Alibaba Cloud is a cloud.

With this important distinction in mind, the battle for bronze (maybe even silver) in the “Cloud Computing Olympics” is no longer so clear cut. A decent chunk of Google Cloud’s 8.9 billion in revenue is probably earned by GSuite, so GCP’s revenue alone is probably closer to Alibaba Cloud’s than at first glance. It’s hard to tell how much of Microsoft’s Commercial Cloud revenue is Azure vs non-Azure, though it’s safe to assume that Azure’s revenue is far bigger than Alibaba Cloud’s revenue. That being said, the growth rate of Azure and Alibaba Cloud are both in the 60-70% range, which could mean two things:

1. Team Azure is wildly out-executing Team Alibaba and Team GCP, even with a larger base to calculate growth rate on (larger the base, typically lower the growth rate over time); or

2. Azure is actually not that much bigger than either GCP or Alibaba Cloud, which may in fact be a reason why Microsoft does not release the unit’s revenue on its own, like AWS or Alibaba Cloud.

Not All Revenue is Created Equal

Based on a projection by IDC, an industry analyst firm, the global public cloud computing spending was $229 billion in 2019 and will grow to $500 billion in 2023. North America will account for more than half of that pie, Western Europe is second with close to 20%. China is pegged as potentially the fastest growing market, but starting from a small base.

What does this mean from a revenue quality standpoint?

Well, it’s clear that the North American market is the most mature, in the sense that enterprises large and small have a solid grasp of what is public cloud computing, its value proposition, and ways to purchase it at scale. Less energy and time need to be spent educating the market; it’s already educated. Thus, revenue coming from this type of customer is of higher quality because the sales cycle is shorter, the tendency to renew for the long term is higher (more predictable revenue growth), and the way customers use different cloud products is more consistent and as intended (less customization, more scale).

In a less mature market, most of these characteristics don’t exist, thus the cost to acquire and keep customers, then scale to multiple customers are all much higher.

It’s old news at this point that it’s hard, if not impossible, for any Chinese tech company to do meaningful business in the U.S. due to political tension, regulatory concerns regarding cybersecurity, and general paranoia about how products with a China link, however tenuous, can lead to security vulnerabilities and intellectual property theft. Alibaba Cloud’s own expansion plan into the U.S. was halted in 2018, after spending more than three years and millions of dollars in building two data centers (one in Northern California, one in Virginia) and a full service American team.

In effect, Alibaba Cloud has been shut out of the most mature and lucrative public cloud market with the highest quality revenue.

On one level, it suggests that Alibaba Cloud’s business is capped. Unless and until the business environment between the U.S. and China improves, it can only play in emerging markets that are less mature.

On another level, you can’t help but be a little impressed by its ability to execute and deliver more than 5 billion USD of run rate primarily from China and other parts of APAC, where the cloud journey is still in its infancy.

When we compare GCP and Alibaba Cloud, we can make a reasonable argument that Alibaba Cloud is executing better and deserves a “third place” nod, because it has grown to be about the same size as GCP, while operating in a much less friendly business environment.

Transparency Is Integrity

Part of the reason why assessing which cloud is bigger/smaller than which is hard is the lack of consistency and transparency in two of the biggest companies’ reporting. This post would’ve been pointless if Microsoft just shared stand alone numbers on Azure and Alphabet did the same for GCP.

Alibaba (and Amazon) deserves some kudos for being straightforward with its reporting.

People usually do not give Chinese tech companies the benefit of the doubt when it comes to transparency and integrity. And those suspicions are generally warranted.

But we should always see things for what they are, not what we tend to believe, and give credit where credit is due. Alibaba Cloud deserves some credit for its transparency.

Opacity breeds suspicion. I’m not insinuating that Microsoft and Alphabet are being convoluted with their cloud reporting for nefarious reasons. But I do hope they will be cleaner in their future reports, and be more like Amazon and Alibaba.

If you like what you've read, please SUBSCRIBE to the Interconnected email list. New posts will be delivered to your inbox (twice per week). Follow and interact with me on: Twitter, LinkedIn.

被遗忘的阿里云


在云计算行业里,被提及最多的公司是所谓的“三巨头”:亚马逊网络服务(Amazon Web Services, AWS)、微软Azure(Azure)、谷歌云平台(Google Cloud Platform,GCP)。

在美国主流商业媒体、华尔街分析师群体里和硅谷媒体中,阿里云通常被忽略。在阿里巴巴最近公布季报的时候,在分析师提出的八个问题中,只有一个提到了阿里云,还是以冠状病毒(COVID-19)为主话题的问题。在硅谷科技媒体The Information最近的一次深度视频讨论(英文)GCP的未来中,只有一位记者不经心的提到一次阿里云。

这种对阿里云的遗漏是有情可原,还是一个盲点?

收益与增长

首先让我们比较一下Azure、GCP和阿里云的收益和增长率。(由于AWS目前是市场上的领头羊,将其纳入这一比较似乎意义不大。)

根据阿里巴巴最新宣布的季报里(2020年2月13日),阿里云此前四个季度的盈利总共创造了51亿美元的收入。前四个季度的同比增长率在62%至76%之间。

Alphabet在最近的一份报告中(2020年2月3日)首次公布了其云计算部门(在财务报表中称为“谷歌云”)的收益,该部门创造了89亿美元。财报并没有分开公布前四个季度里每个季度的业绩,但年同比增长率为53%。Alphabet不单独公布GCP的成绩,只公布“谷歌云”整个的业绩,所以这里也包括其他产品,比如GSuite(例如企业Gmail、域名管理等)。

微软把云的生意在财报里称为“商业云”,在最新宣布的季报里(2020年1月29日)创造了447亿美元的收入。前四季度商业云同比增速在36%-41%之间。Azure过去四个季度的同比增长率在59%至73%之间。与Alphabet一样,微软也不单独公布Azure的成绩(增长率除外),而只是作为整个商业云的一部分。商业云里其实包括杂七杂八一系列东西,如Office 365、Teams、可能还有Skype(因为它是Office生产力产品套餐的一部分)和其他云上订阅版的产品。

447 > 89 > 51。第二,第三,第四。没什么好说的了吧。。。?

并非所有的云都是一样的

可以看到,这几个科技公司报告自己云业务成绩的方式“各有风味”,每个云长得都不一样,完全不统一,使得评估每个公有云平台谁上谁下,非常困难。

然而,对于一个工程师或开发者来说,云的定义很明确。简单明了的说:云就是一组计算、存储和网络链接的资源,加上一些其他常有的基础软件产品和服务(如数据库、容器管理、大数据工具、机器学习框架),可以随时租来用,构建应用程序。

所以,Azure是个云,“商业云”不是。GCP是个云,“谷歌云”不是。阿里云是个云。

考虑到这一重要区别,争夺云计算铜牌(甚至银牌)的战局就不是那么简单明确了。“谷歌云”89亿营收中,相当一部分可能是由GSuite赚取的,因此仅GCP的营收可能和阿里云的差不多。很难确定微软的“商业云”收入中有多少是Azure的,有多少不是,不过可以肯定,Azure的收入还是远远超过阿里云的收入。虽然如此,Azure和阿里云的增长率都在60%-70%之间,这可能意味着两种可能:

  1. Azure团队在商业运维和执行能力上超出阿里和GCP一大截,虽然基数已经很大,但还是能够快速增长(通常基数越大,长期的增长率就会越来越小);
  2. 事实上,Azure也并没有比GCP或阿里云大多少,这可能是微软没有像AWS或阿里云那样单独公布Azure营业额的一个原因。

并非所有的收入都是一样的

根据市场分析所IDC的预测,2019年全球公有云计算支出为2290亿美元,2023年将增长至5000亿美元(英文)。北美市场将占据这块馅饼的一半以上,西欧则以接近20%的份额位居第二。中国被视为可能增长最快的市场,但起点很小。

从收入质量的角度来看,这意味着什么?

很明显,北美市场是最成熟的,从这个层面来看,大大小小的企业对什么是公有云、它的价值是什么,如何大规模购买等基础知识已经熟门熟路了。作为公有云厂商,已经不需要花费精力和时间来教育市场。因此,来自这类客户的收入质量更高,因为销售周期更短,长期续约的欲望更强(便于提高收入的可预测性),云上提供的各种产品的使用方式标准(减低定制个例,便于扩大规模)。

在一个不太成熟的市场中,这些特征大多是不存在的,因此获取和保留客户,再扩展到更多客户的成本都要高得多。

众所周知,中国科技公司在美国做生意很难,甚至完全行不通。原因重重:两国政治和贸易关系的紧张,美国政府对网络安全监管的担忧,以及对与中国有联系的任何产品的普遍偏执,以可能会导致的安全漏洞和知识产权盗窃。阿里云的美国扩张计划已经在2018年叫停(英文)。此前阿里巴巴花了3年多时间和数百万美元建立了两个数据中心(一个在北加州,一个在弗吉尼亚州)和一个完整的美国本土团队,但最后空手而归。

阿里云目前完全不能进入最成熟、利润最丰厚、收入质量最高的美国公有云市场。

从一个层面上讲,这意味着阿里云未来业务的天花板比较低。除非中美之间的商业环境有一天有所改善,否则它只能去不成熟的新兴市场扩展。

从另一个层面看,阿里云团队能在全球商业环境不友好的大势下仍然交付超过50亿美元的收入,主要来自对“上云”还处于起步阶段的中国和亚太市场,还是让人刮目相看的。

比较一下GCP和阿里云,阿里云的执行能力可能更好,应该得到“第三名”的认可,因为它在大的商业环境不利的情况下,已经发展到和GCP差不多的规模。

透明就是诚信

总之,想比比哪个云大,哪个云小,是件让人困扰的事情, 因为行业中的两个巨头的财报既不透明,又不标准。如果微软公布Azure本身的营业额和一些基本常规的数据,Alphabet对GCP也这么做的话,我也就没必要写这篇文章了。

阿里巴巴(和亚马逊)的报告直截了当,值得称赞。

在透明度和诚信方面,人们通常不给中国科技公司加分。这种质疑通常也是有根据的。

但我们应该至始至终以事实说话,而不是听自己脑子里已有的倾向与偏见。谁做的好,做得对,就该加分。阿里云的透明度值得加分。

不透明就会引起怀疑。我并不是在暗示微软和Alphabet不公布自己公有云业绩的背后有什么”邪恶“的理由。我只是希望他们在未来的报告中更直截了当,更像亚马逊和阿里巴巴一样。

如果您喜欢所读的内容,请用email订阅加入“互联”。新文章将会直接送到您的收箱(每周两次)。请在TwitterLinkedIn上给个follow,跟我互动交流!